As the median and paramedian parts of the brain regress with age, the third ventricle slowly begins to widen (59). Yakovlev studied the growth and maturation of the nervous system and noted the regression of the median nuclei of the thalamus and widening of the third ventricle beginning by the fifth decade (59). In addition, he observed a progressive decrease in the size of the massa intermedia (59). In a different study, Morel and Wildi measured the ventricular volume in cadaveric specimens ranging in age from 55 to 99 years and noted an increase in the size of the ventricles up to the ninth decade (59). The ventricles were larger in the men than in the women, and the left lateral ventricle was usually larger than the right (59). In addition to the ventricles becoming larger, Knudson found that the temporal lobes, particularly the hippocampus, uncus, and parahippocampal and fusiform gyri, and area around the insula involute with age (59). The relationship between size of the lateral ventricles and age is more variable than that of the third ventricle (59). In contrast, only mild enlargement of the temporal horns of the lateral ventricles is observed with aging, and the left temporal horn is usually slightly larger than the right.

Multiple brain CT studies have been performed to distinguish normal age-related ventricular changes from pathologic states. Various measurements and ratios have been employed to establish normal metrics, including the Evans, frontal horn, bicaudate, and third ventricle-Sylvian fissure methods (Table 3-1). These studies all found progressive enlargement of the ventricles and cortical sulci (Fig. 3-1) with age. In fact, Nagata et al. reported that the cerebrospinal fluid (CSF)-to-brain ratio remains constant for the first six decades and then begins to change thereafter. One of the most reproducible measures used in these studies is the ratio of ventricular width to the corresponding width of the skull or brain.

Enlargement of the superficial sulci, particularly in the frontal parasagittal region and in the parietal and temporal lobes, is a commonly described normal feature of aging. The etiology for this observation may be related to a decrease in volume of subcortical structures rather than to a change in width of the cortex (59). Miller et al. (61) reported that the ratio of gray matter versus white matter was 1.13 at age 20, 1.28 at age 50, and 1.55 at age 100, which suggests that white matter atrophy exceeds that of gray matter. Most authors suggest that physiologic atrophy commences at the age of 50. However, CT studies on postmortem and living patients do show a wide variation in the CSF space size, ranging between 30% and 50% of the size of healthy young adults. Associated widening of the interhemispheric fissure, extending posteriorly to or beyond the callosomarginal sulci, has also been observed. Progressive widening of the sulci occurs in the frontal lobes and cerebellar vermis, beginning in the teenaged years (13). Widening of the superficial sulci seen with normal aging is often termed cortical atrophy by radiologists; however, in fact, this should more correctly be referred to as gyral or superficial atrophy. Sulcal enlargement in aging is diffuse; however, changes are best identified in specific locations after 50 years of age (12). Widening of the superficial cortical sulci is seen first in the frontal and parietal parasagittal regions (81). The anterior interhemispheric fissure and the CSF spaces around the cerebellar vermis also widen with age (37). The sulci around the central, precentral, postcentral, and superior frontal gyri widen later (49,82).

Studies have shown that white matter volume diminishes by 12% with normal aging and that this age-related change may contribute to the development of mild cognitive impairment (MCI) in the aged. MCI is now being proposed as a transitional stage between normal aging and dementia (30). Studies have shown that 30% to 80% of elderly individuals without neurologic deficits have focal abnormalities in the white matter (11). These foci are seen as areas of high signal in the periventricular, subcortical white matter, with capping of the lateral ventricular margins on T2-weighted images. On electron microscopy, these corresponding MRI changes feature atrophy of axons and myelin with associated gliosis, tortuous thickened vessels, and increased intracellular water (23). Some of these changes histologically are thought to be secondary to ectasia of the arterioles, with enlargement of the surrounding perivascular spaces (23). These findings were first described by Durand-Fardel in 1843 (28) and were termed état criblé. In addition, aging is associated with an increase in iron deposition specifically in the corpus striatum (2). This iron accumulation is thought to be related to a decrease in oligodendroglial function and dopamine production and an increase in the free radical formation (2). Iron is also deposited in the walls of blood vessels. Iron is more easily visualized on T2-weighted images as focal regions of low signal secondary to field heterogeneity and magnetic susceptibility (25).

Finby and Kraft (59) found that most skull radiographs taken of the same individuals over time revealed an increase in the size of the cranial vault, facial bones, and paranasal sinuses and in the skull thickness, which is thought to be secondary to continuous resorption and regrowth of the skull and facial bones.

Magnetic resonance spectroscopy (MRS) is an MRI technique that permits noninvasive quantitative measurement of cerebral biochemical metabolites. MRS can be used to selectively “sample” a biochemical map of a selected volume of the brain (voxel) (77). The principle cerebral metabolites that are routinely measured with MRS are N-acetylaspartate (NAA), creatine and phosphocreatine (PCr/Cr), choline (Cho), myoinositol (mI), and glutamate plus glutamine (Glx). NAA is principally found within neurons and is a sensitive indicator of neuronal loss. Creatine is a byproduct of oxidative phosphorylation and the energy cycle associated with adenosine triphosphate (ATP) production; the concentration of this metabolite is very constant, and therefore, this value is often used as an internal reference standard. Myoinositol is a byproduct of glucose metabolism and is known to occur in elevated concentrations in diabetes and Alzheimer’s disease. Choline (Cho) is abundant in the cell membrane, and elevation in this metabolite’s concentration is reported in neoplasia and demyelination (72).

As seen in Table 3-2, the normal aging brain shows very small but definite changes in the MRS spectral signature as compared with a young adult. With aging, NAA has been shown to decrease in the occipital gray matter, the generalized concentration of choline is reported to elevate, and myoinositol (mI) is relatively reduced. Ross et al. (72) has shown that none of these changes exceed 10% of the normal young adult.

The complex but close relationship between brain physiologic activity and brain blood flow is the basis for nuclear medicine imaging protocols for the brain. The development of scintillation multiprobe systems provided a technique to quantify regional cerebral blood flow (rCBF), or perfusion, within individual regions of the cortex and, ultimately, to compare the rCBF with regional brain function. Single photon emission computed tomography (SPECT) and positron emission tomography (PET) have the ability to perform rapid three-dimensional physiologic imaging of the brain. This has become particularly important in geriatric brain imaging by helping to differentiate the various causes of dementia. For example, the reduction of metabolism and perfusion in the temporoparietal areas in the brain has become a biomarker for Alzheimer’s disease (AD). Another potential use for PET biomarkers is to follow the response to drug or other treatments (39).

The radiopharmaceuticals that are in common use (1) to evaluate brain physiology by rCBF SPECT include ^99mTc-hexamethyl propylene amine oxime (HMPAO; exametazime, Ceretec, Amersham Inc.), ¹²³I-inosine monophosphate (IMP; which is intermittently available), ¹³³Xe gas (General Electric Corp.), and ^99mTc-ethyl cysteinate (ECD; Merck Du Pont Inc.).

The most commonly used radiopharmaceuticals for PET imaging (22) of the brain are ¹⁸F-fluorodeoxyflucose (18F-FDG), which measures the cerebral metabolic rate for glucose (CMRGlc), and ¹⁵O-H₂O, which measures rCBF. Two new compounds being used for imaging in the AD patient are fluorine-18-labelled FDDNP and the Pittsburgh compound B (carbon-11-labelled PIB). The [¹⁸F]FDDNP labels both neurofibrillary tangles and beta-amyloid neuritis plaques, whereas [¹¹C]PIB only labels the amyloid deposits (83).

Normal age-related atrophy can significantly influence qualitative and quantitative analyses of 18FFDG-PET studies. The 18F-FDG-PET findings in a normally aging brain, as reported in the literature, have been inconsistent. A number of investigators have described diminished regional glucose metabolism (CMRGlc) in the temporal, parietal, somatosensory, and, especially, frontal regions. Others have described more prominent decreases in the frontal and somatosensory cortices in comparison with younger controls (15,51,87). When brain atrophy was not considered, mean CMRGlc values were lower in older patients, particularly in the frontal, parietal, and temporal regions. Also, women had significantly higher mean CMRGlc than men. Cerebrovascular risk factors in the population were also seen not to have any effect on CMRGlc (87).

Volumetric MRI is a relatively new and increasingly important technique in the study of dementias, especially in the aid of diagnosis and to follow progression of disease (85). The availability of inexpensive powerful computing platforms provides methods for automatic volumetric analysis of three-dimensional datasets. Voxel-based morphometry involves a voxel-wise statistical comparison of the local concentration of gray matter between two groups of subjects. It is used to assess patterns of atrophy (85).

With more people living longer, the prevalence of AD has doubled since 1980 and will likely triple by the year 2050 (14). Although it is possible to make an accurate clinical diagnosis of dementia in most patients with severe disease, it is very difficult to differentiate between AD and other dementing disorders in patients with mild disease (80). With functional imaging studies such as SPECT and PET, it is believed to be possible to make an early diagnosis of AD and possibly help in elucidating the mechanisms underlying the disorder. AD is known to target specific brain regions, especially the cholinergic basal forebrain and medial temporal lobe structures including the hippocampus, amygdala, and entorhinal cortex (14). The role of neuroimaging has been detection of early structural changes in the hippocampal formation and parahippocampal gyrus because memory loss is a prominent feature of AD. Table 3-3 summarizes the benefits and disadvantages of various imaging techniques.